Fixed Code

temp.py

import numpy as np
import bson
import tensorflow as tf
from StringIO import StringIO
import skimage
from skimage import io


# Load bson data and dump to TFRecord File
def _int64_feature(value):
    return tf.train.Feature(int64_list=tf.train.Int64List(value=[value]))


def _bytes_feature(value):
    return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))


opts = tf.python_io.TFRecordOptions(tf.python_io.TFRecordCompressionType.ZLIB)

train_filename = 'train.tfrecords'
writer = tf.python_io.TFRecordWriter(train_filename, options=opts)

data = bson.decode_file_iter(open('/home/hadi/Downloads/train_example.bson', 'rb'))
for c, d in enumerate(data):
    product_id = d['_id']
    category_id = d['category_id']
    for e, pic in enumerate(d['imgs']):  # loop through each picture
        img_raw = pic['picture']
        img_raw = skimage.io.imread(StringIO(img_raw)).astype(np.uint8).tostring()
        label = category_id
        feature = {'label': _int64_feature(label),
                   'image': _bytes_feature(img_raw)}
        # Create an example protocol buffer
        example = tf.train.Example(features=tf.train.Features(feature=feature))

        # Serialize to string and write on the file
        writer.write(example.SerializeToString())
writer.close()

import matplotlib.pyplot as plt

data_path = 'train.tfrecords'

with tf.Session() as sess:
    filename_queue = tf.train.string_input_producer([data_path], num_epochs=1)

    reader = tf.TFRecordReader(options=opts)
    _, serialized_example = reader.read(filename_queue)

    features = tf.parse_single_example(
        serialized_example,
        features={
            'label': tf.VarLenFeature(tf.int64),
            'image': tf.FixedLenFeature([], tf.string)
        })

    image = tf.decode_raw(features['image'], tf.uint8)
    label = tf.cast(features['label'], tf.int32)
    image = tf.reshape(image, [180, 180, 3])
    images, labels = tf.train.shuffle_batch([image, label], batch_size=10, capacity=30, num_threads=1,
                                            min_after_dequeue=10)

    # Initialize all global and local variables
    init_op = tf.group(tf.global_variables_initializer(), tf.local_variables_initializer())
    sess.run(init_op)

    # Create a coordinator and run all QueueRunner object
    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(coord=coord)

    for batch_index in range(5):
        img, lbl = sess.run([images, labels])
        img = img.astype(np.uint8)
        print('Batch %d and Batch shape is %s' % (batch_index + 1, img.shape))
        for j in range(10):
            plt.subplot(2, 5, j + 1)
            plt.imshow(img[j, ...])
        plt.show()
        # Stop the threads
    coord.request_stop()

    # Wait for threads to stop
    coord.join(threads)
    sess.close()